summaryrefslogtreecommitdiff
path: root/sys
Commit message (Collapse)AuthorAgeFilesLines
* Move up to 5.2-RELEASE!release/5.2.0Scott Long2004-01-101-1/+1
| | | | | Notes: svn path=/releng/5.2/; revision=124329 svn path=/release/5.2.0/; revision=124330; tag=release/5.2.0
* MFC rev 1.9:Scott Long2004-01-101-0/+1
| | | | | | | | | Disable the APIC when selection the 'Safe Boot' option. Approved by: rwatson Notes: svn path=/releng/5.2/; revision=124325
* MFC: Make ip6fw reset work for TCP packets with SYN set.David Malone2004-01-091-0/+2
| | | | | | | Approved by: re, sam, ume Notes: svn path=/releng/5.2/; revision=124320
* MFC 1.138 - Check for transmit status and free the mbuf chain whenDon Lewis2004-01-092-3/+5
| | | | | | | | | | | | | | processing the last transmit descriptor for the frame instead of the first descriptor. This eliminates the bug that caused the driver to report spurious collisions and output errors that was introduced when the driver was changed to look at the first transmit descriptor when it was converted to use busdma (rev 1.115). Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=124315
* MFC: Limiters and sanity checks for TCP MSS resource exhaustion attacks.Andre Oppermann2004-01-091-0/+24
| | | | | | | | | | The net.inet.tcp.minmssoverload is set to zero and thus connection drop is disabled by default. Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=124295
* MFC rev 1.20: Reduce TCP_MINMSS default to 216.Andre Oppermann2004-01-091-4/+3
| | | | | | | Approved by: re (scottl, implicit) Notes: svn path=/releng/5.2/; revision=124291
* MFC: Limiters and sanity checks for TCP MSS resource exhaustion attacks.Andre Oppermann2004-01-095-4/+92
| | | | | | | | | | The net.inet.tcp.minmssoverload is set to zero and thus connection drop is disabled by default. Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=124286
* MFC rev 1.173: If path mtu discovery is enabled set the DF bit in allAndre Oppermann2004-01-091-0/+4
| | | | | | | | | | cases we send packets on a tcp connection. PR: kern/60889 Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=124285
* MFC rev 1.204: Do not set the ip_id to zero when DF is set on packet.Andre Oppermann2004-01-091-12/+6
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=124284
* MFC 1.92: remove extraneous unlock that can lead to a panicSam Leffler2004-01-081-1/+1
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=124254
* MFC: Copyright 2004Warner Losh2004-01-081-3/+3
| | | | | | | approved by: re@ (scottl) Notes: svn path=/releng/5.2/; revision=124239
* Merge critical bugfix (panic due to NULL pointer deref) from rev 1.30.Joerg Wunsch2004-01-071-1/+4
| | | | | | | Approved by: re Notes: svn path=/releng/5.2/; revision=124218
* MFC: Rev. 1.8Martin Blapp2004-01-061-1/+1
| | | | | | | | | | | | The reset_type should not be 0x80, it should be set to zero. 0x80 can cause the command to be rejected as invalid. This bug exists also in the Linux IPS ffdc code. Submitted by: David Jeffery (Adaptec) Approved by: scottl (re) Notes: svn path=/releng/5.2/; revision=124170
* Plug file descriptor leak, as discussed on -current.David Malone2003-12-301-0/+1
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=124004
* MFC:Poul-Henning Kamp2003-12-304-4/+15
| | | | | | | | | | | per provider withering (for atapi-cd) G_PF_ORPHAN flag & check. (for atapi-cd) mtx_init NULL/MTX_DEF confusion. Approved by: re@ Notes: svn path=/releng/5.2/; revision=123993
* MFC: Improvements to if_sis driver.Poul-Henning Kamp2003-12-302-17/+31
| | | | | | | | | | Don't spend interrupt time stopping already stopped interfaces. Fix short-cable handling code. Approved by: re@ Notes: svn path=/releng/5.2/; revision=123991
* MFC: Fix acpi_MatchHid() to check compatability IDs.John Baldwin2003-12-291-3/+3
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123973
* MFC 1.97: fix double mutex initialization when statically configuredSam Leffler2003-12-231-11/+27
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123782
* MFC rev. 1.52: IN_MULTICAST wants an address in host byte order.Maxim Konovalov2003-12-231-1/+1
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123763
* MFC Rev 1.71SUZUKI Shinsuke2003-12-221-1/+1
| | | | | | | | | fixed an IPv6 path MTU discovery failure owing to a lack of initialization. approved by: re(scottl) Notes: svn path=/releng/5.2/; revision=123712
* MFC: Add support for _CID package types and use dynamic storage forNate Lawson2003-12-193-62/+123
| | | | | | | | | | | all of our devinfo structs. Revisions: acpi.c:1.110; acpi_pcib.c:1.35,1.36; acpi_pci_link.c:1.11 Approved by: re (scottl, in a new record time) Notes: svn path=/releng/5.2/; revision=123654
* MFC: Sync ata driver with head.John Baldwin2003-12-185-12/+40
| | | | | | | | | | | - Various fixes for SiI3112A. - Support for the SiS 661 chipset. - Fix register timings for AMD/VIA/nVidia chipsets. Approved by: scottl (re) Notes: svn path=/releng/5.2/; revision=123652
* MFC: Handle empty channels in earlier fix for missing slaves.John Baldwin2003-12-181-3/+7
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123651
* - MFC 1.475, return EBUSY if NOWAIT was passed to vget() and the XLOCK wasJeff Roberson2003-12-181-3/+4
| | | | | | | | | held. Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123639
* - MFC 1.474, properly cleanup VCHR after forced unmount.Jeff Roberson2003-12-181-1/+28
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123638
* MFC: add sys/_null.h and change the definition of NULL to a long on ia64Peter Wemm2003-12-183-9/+4
| | | | | | | | | | | and amd64. Ensure that this is not accidently activated on any platforms other than these two since its quite late in the release cycle and it is not worth the risk of disturbing to the established platforms. Approved by: re (scottl, jhb) Notes: svn path=/releng/5.2/; revision=123618
* MFC: make preloaded filesystems work with md(4) again.Poul-Henning Kamp2003-12-171-1/+1
| | | | | | | Approved by: re@(scottl) Notes: svn path=/releng/5.2/; revision=123613
* MFC: Make msdosfs long filenames matching case insensitive again.Tim J. Robbins2003-12-173-11/+19
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123598
* MFC rev. 1.33: Don't sign-extend characters in file names.Tim J. Robbins2003-12-171-2/+2
| | | | | | | Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123590
* Don't free(9) modified pointer.Poul-Henning Kamp2003-12-161-11/+17
| | | | | | | Approved by: re@ Notes: svn path=/releng/5.2/; revision=123575
* This commit was manufactured by cvs2svn to create branch 'RELENG_5_2'.cvs2svn2003-12-154-0/+4181
| | | | Notes: svn path=/releng/5.2/; revision=123545
* MFC: 1.13: use the correct symbol for declaring the sysvec table sizesPeter Wemm2003-12-131-1/+2
| | | | | | | Approved by: re (jhb, scottl) Notes: svn path=/releng/5.2/; revision=123524
* RegenPeter Wemm2003-12-134-17/+25
| | | | | | | Approved by: re (scottl, jhb) Notes: svn path=/releng/5.2/; revision=123523
* MFC: 1.27-1.28: sync up syscalls listPeter Wemm2003-12-131-9/+8
| | | | Notes: svn path=/releng/5.2/; revision=123522
* MFC 1.226 - Set fs_ronly to the correct value in ffs_reload() whenDon Lewis2003-12-121-0/+2
| | | | | | | | | | reloading the file system super block after fsck has repaired the file system. Approved by: re (scottl + jhb) Notes: svn path=/releng/5.2/; revision=123482
* MFC revision 1.175Alexander Kabaev2003-12-111-2/+6
| | | | | | | | | | Addendum to revision 1.174: In the case where vm_pager_allocate() is called to create a vnode-backed object, the vnode lock must be held by the caller. Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123469
* Up the timeout on a command from 500ms to 5s unconditionally. ThisWarner Losh2003-12-111-1/+1
| | | | | | | | | | | | | | | | will up the failure time from about 1s to about 7s in the worst case that I can think of. Other fixes have brought that number down from 30 minutes, so I don't feel too bad about it. Some older versions of the Prism cards have issues that make them unable to respond in 500ms for some commands. Current has a more extensive fix that only penalizes the old firmware, but this is a more conservative change more appropriate to a frozen branch. Suggested/Submitted by: iedowse Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123467
* MFC: Replace fix for empty turnstile race with a more general fix for theJohn Baldwin2003-12-111-10/+12
| | | | | | | | | | | more general race of a thread that was already woken up but still pending. Adjust assertion for TDF_TSNOBLOCK case to allow for a thread to be on the run queue as well as currently running. Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123462
* Nursemaid: Fix compilation.Bruce M Simpson2003-12-101-4/+4
| | | | | | | Sponsored by: Weyland-Yutani Corporation Notes: svn path=/releng/5.2/; revision=123387
* Don't build a kernel.debug for the release.Scott Long2003-12-075-5/+5
| | | | Notes: svn path=/releng/5.2/; revision=123274
* MFC: 1.85: fix the runq selection when load balancing runs on machines withPeter Wemm2003-12-071-1/+1
| | | | | | | | | 64 bit runq status bit fields. Approved by: re (scottl) Notes: svn path=/releng/5.2/; revision=123251
* MFC: Don't call em_stop() when the watchdog expires. This preventsDaniel Eischen2003-12-071-1/+0
| | | | | | | | | a panic caused by a mutex assertion. Approved by: re@(scottl) Notes: svn path=/releng/5.2/; revision=123241
* Disable WITNESS, INVARIANTS, and DDB for our Tier-1 platforms for the releaseScott Long2003-12-075-19/+19
| | | | Notes: svn path=/releng/5.2/; revision=123205
* RELENG_5_2 is at 502000Scott Long2003-12-071-1/+1
| | | | Notes: svn path=/releng/5.2/; revision=123198
* RELENG_5_2 is at 5.2-RC nowScott Long2003-12-071-1/+1
| | | | Notes: svn path=/releng/5.2/; revision=123197
* This commit was manufactured by cvs2svn to create branch 'RELENG_5_2'.cvs2svn2003-12-07277-180865/+0
| | | | Notes: svn path=/releng/5.2/; revision=123193
* Reconfigure the runq macros to use the 64 bit ffs/bsf routines insteadPeter Wemm2003-12-061-5/+5
| | | | | | | | | | of doing a loop and taking two 32 bit passes at the runqueue bits. All the 64 bit platforms should probably do this since there are 64 run queues. Approved by: re (scottl) Notes: svn path=/head/; revision=123182
* Add 64 bit bsf*/ffs* routines. Have the ffs() inline use gcc's builtinPeter Wemm2003-12-061-1/+40
| | | | | | | because it uses the better cmove instructions to avoid branches. Notes: svn path=/head/; revision=123181
* Various whitespace and cosmetic sync-up's with i386.Peter Wemm2003-12-069-12/+13
| | | | | | | Approved by: re (scottl) Notes: svn path=/head/; revision=123180
* amd64_protection_init and the protection_codes[] array was overkill.Peter Wemm2003-12-061-50/+8
| | | | | | | | | Inline it instead. Approved by: re (scottl) Notes: svn path=/head/; revision=123179
generated by cgit v1.3 (git 2.53.0) at 2026-05-09 02:27:48 +0000